Jilin Chen
Zerozero88 the News Recommender
(Click the image to enlarge)

Twitter streams are on overload, and Twitter users struggle to keep on top of constant news updates from their social network. In this research we designed zerozero88, a web application that recommends news URLs to Twitter users. Zerozero88 achieves personalization by inferring topic interest and influence of people in each user's social network and making recommendations accordingly.

Implemented in a highly scalable Hadoop MapReduce based infrastructure, zerozero88 runs across a cluster of 24 machines and updates its recommendation hourly to reflect the most recent updates on Twitter. Zerozero88 constantly collects, stores and analyzes Twitter messages from hundreds of thousands Twitter users.

Chen, J., Nairn, R., Nelson, L., Bernstein, M., and Chi, E.H. Short and Tweet: Experiments on Recommending Content from Information Streams. CHI 2010. [PDF]
Tenure and Interest Diversity in Wikipedia Work Groups

Social psychology and management science has demonstrated the profound effects of diversity in offline work groups. In this research we quantitatively analyzed the activities of 20,000+ Wikipedia editors in 600+ work groups over three and half years, and addressed the question of how tenure and interest diversity among group members affects the performance of groups.

The analysis is guided by prior theories and extensive studies on offline work groups, and performed with Hierarchical Linear Modeling, an advanced form of linear regression capable of modeling temporal and intra group correlation in the activity data.

Chen, J., Ren, Y., and Riedl, J. The Effects of Diversity on Group Productivity and Member Withdrawal in Online Volunteer Groups. CHI 2010. [PDF]
People Recommender in Enterprise

Many enterprises have set up internal social network sites for their employees, and employees are often interested in expanding their network on these sites. In this research we studied a people recommender on SocialBlue (formerly Beehive), an IBM internal social network site. The recommender was designed to help IBM employees find known, offline contacts and discover new friends. We deployed the recommender alive and conducted an online evaluation involving 3,500 users. The recommender was found effective in expanding users' friend lists.

Chen, J., Geyer, W., Dugan, C., Muller, M., and Guy, I. "Make New Friends, but Keep the Old" - Recommending People on Social Networking Sites. CHI 2009. [PDF]

The views and opinions expressed in this page are strictly those of the page author.
The contents of this page have not been reviewed or approved by the University of Minnesota.